How FtsZ-rings get shov­eled to mid­cell in vitro

Pic­tures Con­sid­ered #61

by Christoph

The last thirty years have been an "eye-open­ing" jour­ney for mi­cro­bial cell bi­ol­o­gists, quite lit­er­ally. I am talk­ing, of course, about flu­o­res­cent la­bel­ing of in­di­vid­ual cell com­po­nents and the non‑invasive obser­va­tion of their "be­hav­ior" in liv­ing bac­te­r­ial cells. Just re­cently, I con­sid­ered a video clip about Vib­rio cholerae har­poon­ing DNA. Now again a video clip, this time about In vitro as­sem­bly, po­si­tion­ing and con­trac­tion of a di­vi­sion ring in min­i­mal cells − the no­to­ri­ous FtsZ ring − as the ti­tle of a re­cent pa­per by Ko­hyama et al. (2022) goes.

Pe­tra Schwille's lab at the MPI für Bio­chemie in Mar­tin­sried, Ger­many, has set it­self the am­bi­tious goal of re­ca­pit­u­lat­ing the com­plex process of cell di­vi­sion in E. coli by its re­con­struction in vitro. Their most re­cent pa­per de­scribes the cur­rent achieve­ment: func­tional es­tab­lish­ment of the MinCDE os­cil­la­tor in an ar­ti­fi­cial cell, a mem­brane-bound vesi­cle, that "shov­els" newly syn­the­sized FtsZ fil­a­ments and "ringlets" to the equa­tor of the vesi­cle that then shows signs of con­stric­tion. All this within a lit­tle more than two hours of incu­bation, and sta­ble over this time span (Fig­ure 1 here is Sup­ple­men­tary Movie 8 of their pa­per).


Fig­ure 1. Zoom-in view of a vesi­cle that formed FtsZ-ring dri­ven by Min waves and con­stric­tion of the vesi­cle by FtsZ-ring. 3D max pro­jec­tion of FtsZ-G55- Venus-Q56 and mCherry-MinC chan­nel (Left panel). Ran­domly dis­tributed FtsZ mesh was then re­or­ga­nized into ring struc­ture by the pole-to-pole os­cil­la­tion of Min waves. Af­ter for­ma­tion of FtsZ-ring, Min waves sta­bly kept pole-to-pole os­cil­la­tion over 1 h, and FtsZ-ring constrict­ed the vesi­cle. (Right panel) Equa­to­r­ial plane of the vesi­cle vi­su­al­ized by Dif­fer­en­tial in­ter­fer­ence con­trast. Im­ages clearly show the "neck" of the con­stricted ve­sicle by FtsZ-ring. The vesi­cle it­self slightly moves back and forth syn­chron­i­cally to the pole-to-pole os­cil­la­tion of Min waves. Record­ing was started af­ter 20 min in­cu­ba­tion at 37°C. Time in­di­cates hh:mm:ss. Scale bar: 10 μm. Sup­ple­men­tary Movie 8 of Source

What you see   in the video clip, side by side, are the si­mul­ta­ne­ously filmed im­ages of a ve­sicle in the light mi­cro­scope (right) and the flu­o­res­cence ex­cited by laser light in­side this ve­sicle (left). The two E. coli pro­teins that you see mov­ing within the vesi­cle are FtsZ in "green" and MinC in "ma­genta" (the cap­tion to Fig­ure 1 gives more de­tails). Pay at­ten­tion to the ti­mer, which tells you that 2 hours and 15 min­utes of ob­ser­va­tion are con­densed into the time-lapse video of 40 sec­onds that you see in loop. The dy­namic processes thus run con­si­der­ab­ly slower in real time! Also note the scale bar, which tells you that the di­am­e­ter of the vesi­cle is ~40 µm, that is, roughly 10‑times that of an E. coli cell.

At 0:06 (h:min) into the video, you can see faintly shin­ing "green" FtsZ fil­a­ments dis­trib­uted ran­domly on the in­ner side of the vesi­cle mem­brane that be­come more promi­nently vis­i­ble from 0:16 on­wards. At around 0:25 you can see the spe­cific growth mode of FtsZ fil­a­ments called "thread­milling": in a GTP-hy­drol­y­sis-de­pen­dent re­ac­tion, FtsZ monomers add to one end of the fil­a­ment, while monomers de­tach from the other end. Growth or shrink­age of the fil­a­ment is then de­ter­mined by vary­ing on/off rates (see here for a di­a­gram).

Af­ter 0:16, you also see a coarse mesh­work of FtsZ "ringlets" of vary­ing di­am­e­ters con­co­mit­ant with a de­crease in the num­ber of fil­a­ments. Then, af­ter 0:46, you see the vesi­cle wob­ble and a faintly shin­ing cloud of "ma­genta" MinC. From 1:04 on­wards you can see that the MinC pro­tein os­cil­lates be­tween both halves of the vesi­cle, but the com­pres­sion of the time-lapse pre­vents you from see­ing the ac­tual am­pli­tude of the os­cil­la­tion, which is close to 20 sec (see Fig­ure 3, Bot­tom part). Keep in mind that you see a two-di­men­sional im­age of a 3D ob­ject. Thus, what ap­pears as a "cloud" of MinC is ac­tu­ally the lin­ing of the in­side of the ve­sicle mem­brane by MinC rather than its dif­fuse dis­tri­b­u­tion in the "cy­to­plasm" as ex­plained be­low. In sev­eral in­stances you see in ad­di­tion to the "cloud" a pro­nounced ac­cu­mu­la­tion of MinC in "po­lar" re­gions of the vesi­cle.

Fig­ure 2. Re­con­sti­tu­tion of min­i­mal di­vi­sion ring pla­ce­ment sys­tem by co­ex­pres­sion of five es­sen­tial com­ponents. a Schematic il­lus­tra­tion of co­ex­pres­sion sys­tems of FtsA, FtsZ-G55-Venus-Q56 (green), mCher-ry-MinC (ma­genta), MinD, and MinE. b 3Dmax pro­jec­tion of FtsZ-ring for­ma­tion and Min pro­tein lo­cal­iza­tion at the pole of the vesi­cles af­ter 100 min of co-ex­pres­sion at 37 °C (1 nM ft­sAopt, 3 nM ft­sZopt-G55-Venus-Q56, 2 nM mCherry-MinC, and 1 nM minDE operon tem­plates). mCherry-MinC and FtsZ-G55-Venus-Q56 are in­di­cated in ma­genta and green, re­spec­tively. Scale bar: 20 μm. A white as­ter­isk marks the li­po­some fea­tured in Fig. 1. Source

Also af­ter 1:04, most of the FtsZ "ringlets" grow to the di­am­e­ter of the vesi­cle, and their flu­o­res­­cence be­comes brighter. This in­di­cates a bundling of length­wise aligned FtsZ fi­­la­ments that fi­nally re­sults in the large "rings," all the while the in­di­vid­ual fil­a­ments still thread­mill. From 1:21 to the end of the video you see in the phase-con­trast light mi­croscopy (right pa­nel) a slight in­dent, a con­stric­tion, that cor­re­sponds to the re­gion where FtsZ rings suc­ces­si­vely accumu­late (left panel).

An aside: you surely no­ticed the tiny green­ish dot on the right side be­low the large vesi­cle. This is a tiny vesi­cle that you also see in Fig­ure 2b, which shows that vesi­cle sizes in their experi­ments var­ied a lot.

What you can not see   but need to know to ap­pre­ci­ate the movie. The cell-free ex­pres­sion sys­tem of Ko­hyama et al. (2022) pro­vided "in­vis­i­ble," that is un­tagged FtsA pro­tein in ad­di­tion to tagged, "green" FtsZ (see Fig­ure 2a). Mem­brane-bound FtsA is in­stru­men­tal in tran­siently at­tach­ing FtsZ fil­a­ments to the mem­brane, and with­out it more straight FtsZ fi­la­ments would have been ran­domly dis­trib­uted in the vesicle's lu­men. The var­i­ous large FtsZ rings seen, for ex­am­ple at 1:27:20, would co­a­lesce into one sin­gle com­pact ring as in real cells if ZapA were present. E. coli ZapA is known to co­op­er­a­tively in­crease the spa­tial or­der of the fil­a­ment net­work, but binds only tran­siently to FtsZ as sort of a crosslinker. How­ever, ZapA was not in­cluded in the cell-free ex­pres­sion sys­tem, and Ko­hyama et al. (2022) ex­plain that so far they could not ob­tain func­tional ex­pression sys­tems for not more than the 5 com­po­nents in­di­cated in Fig­ure 2a.

You can­not "see" in the movie the un­tagged MinD and MinE pro­teins that were in­cluded in the ex­pres­sion sys­tem to­gether with "ma­genta" tagged MinC (Fig­ure 2a). To ex­plain this bet­ter I give a brief in­tro­duc­tion to the MinCDE sys­tem.

Fig­ure 3. Top: Plas­mid pDR112[Plac::minDE-gfp] was in­tro­duced into strains PB103 [wt] (a–c), and cells were grown in the pres­ence of 25 µM IPTG and fixed be­fore ob­ser­va­tion by flu­o­res­cence (a–c) and DIC (a'–c') mi­croscopy. The bar rep­re­sents 2.0 µm. Source. Bot­tom: (H) Time-lapse im­ages show­ing Gfp-MinD os­cil­la­tion in normal­ly di­vid­ing cells of strain PB103(λDR122) [wt(Plac:: gfp-minDE)]; Times are in­di­cated in sec. (H') DIC, cell length ~2 µm (acc. to scale bar in other part of the ori­ginal fig­ure). Source

The MinCDE os­cil­la­tor   is a fas­ci­nat­ing tran­sient dy­namic struc­ture in grow­ing E. coli cells. Guided by MinD, MinE forms a ring-like struc­ture, the MinE ring, along the in­ner mem­brane at mid­cell, in the re­gion of later con­stric­tion, sep­ta­tion, and cell di­vi­sion (Fi­gure 3, Top). MinE can be vi­su­al­ized in elon­gat­ing cells up to the point where the weakly in­dented mid­cell re­gion be­comes a promi­nent con­stric­tion, in­di­cat­ing its dis­as­sem­bly (Fi­gure 3, com­pare b/b' with c/c'). For­ma­tion of the MinE ring is in­de­pen­dent of FtsZ ring for­mation, and FtsZ and MinE rings ap­pear rather to abut and not su­per­im­pose by mi­croscopy. MinD binds to the in­ner mem­brane and forms large oligomeric patches to­gether with MinC ("ma­genta" MinC in the Fig­ure 1 movie). These patches form and dis­solve in an os­cil­lat­ing man­ner in both cell halves depen­dent on their lo­cal con­cen­tra­tions (Fig­ure 3, Bot­tom). MinD binds to both MinE and MinC but not si­mul­ta­ne­ously. This is what gets the os­cil­la­tor started, and leads to MinE ac­cu­mu­la­tion as ring-like struc­ture at mid­cell while MinC ac­cu­mu­lates in both po­lar re­gions (see here for a more de­tailed de­scrip­tion of the os­cil­la­tor).

It was al­ready found in 1993 that FtsZ‑ring for­ma­tion at mid­cell is sup­pressed in a Δmin mu­tant (FtsZ lo­cal­izes ran­domly and even­tu­ally forms rings at sites close to the cell poles). Thus, MinCD patches al­ter­nat­ing be­tween both cell halves pre­vent the FtsA-me­di­ated at­tachment of FtsZ fil­a­ments and re­sult in their ac­cu­mu­la­tion as FtsZ rings at mid­cell. This is an al­most bra­zen but ap­prox­i­mately cor­rect sum­mary of the cur­rent state of knowl­edge. And it is what you see in the video (Fig­ure 1).

To con­clude, I have to come back to a quip⛓️‍💥 that Emanuele Sev­eri had heard at a con­fer­ence and shared on Twit­ter: "What is true for E. coli might be true for an ele­phant, but cer­tainly not for Bacil­lus!" I know that Erin Go­ley and Yves Brun would add: "...and even less for Cau­lo­bac­ter." Yes, it's true, FtsZ rings form at mid­cell at the on­set of cell di­vi­sion in both Bacil­lus and Caulobac­ter. Also most other coc­coid or rod-shaped bac­te­ria across the nu­mer­ous phy­la have ftsZ genes; even sym­bionts with re­duced genomes like Buch­n­era aphidi­cola (0.65 Mb) and Stam­mera capleta (0.27 Mb) have them (if and how they form FtsZ rings dur­ing cell di­vi­sion is everyone's guess). But both Bacil­lus and Caulobac­ter lack the MinCDE os­cil­la­tor and each have com­pletely dif­fer­ent mech­a­nisms for po­si­tion­ing their Z rings at mid­cell. This is the is­sue with vir­tu­ally all bac­te­r­ial "model sys­tems" for cel­lu­lar processes in "model or­ganisms" like E. coli: close and dis­tant rel­a­tives tend to use very sim­i­lar vo­cab­u­lary but em­ploy of­ten dras­ti­cally dif­fer­ent syn­tax and gram­mar.

Fig­ure 4. Pro­gram book­let, Chorin work­shop 1997. By the au­thor

My per­sonal at­trac­tion for mi­croscopy movies goes back to a work­shop in 1997 (Fig­ure 4). We had brought to­gether al­most all the lead­ing re­searchers work­ing on the bac­te­r­ial cell cy­cle at that time − many still do to­day. To name just a few "di­vi­sion­ists," Joe Lutken­haus talked about the FtsZ ring, of course (see Pic­tures Con­sid­ered #5), Yves Brun and James Gober about FtsZ in Caulobac­ter, Miguel Vi­cente and David Bramhill about FtsZ in E. coli, Keith Chater about cell sep­ta­tion in Strep­to­myces (see Pic­tures Con­sid­ered #52), Jeff Er­rington and Alan Gross­man about chro­mo­some seg­re­ga­tion in veg­e­ta­tive and sporu­lat­ing Bacil­lus (see Sub­tle Bacil­lus sub­tilis in STC). And then there was Piet de Boer in­tro­duc­ing the at­ten­dants to the MinE ring and the MinCDE os­cil­la­tor in E. coli. Younger read­ers won't know: we didn't have reg­u­lar Pow­er­Point pre­sen­ta­tions back then. Pre­sen­ters had their man­u­script and a stack of slides, hope­fully shown in the right or­der from a man­u­ally op­er­ated pro­jec­tor ("next slide please"). At the end of his talk, Piet said rather ca­su­ally that he could show to those in­ter­ested a short movie on his lap­top about the os­cil­la­tion of MinC in liv­ing cells dur­ing the lunch break. Gosh! I could hardly take my eyes off Piet's screen, and, de­spite the rather poor qual­ity of the video clip, Yves Brun and I con­firmed to each other that this was a jaw‑dropping mo­ment. I vividly re­mem­bered this episode when, twenty-five years later, I first saw the tech­ni­cally bril­liant, beau­ti­ful MinC os­cil­la­tions in the movie con­sidered here.

 

Do you want to com­ment on this post? We would be happy about it! Please com­ment on Mastodon or on Bluesky.
 

Other Posts